Alarm/Event based notification

Brashquido · Post by **Brashquido** » Mon Mar 01, 2004 12:00 am

Hi All,

First off a HUGE congratulations to the Cacti dev team for making such a great product available to us all, it is greatly appreciated. I've been in IT 8 years, and I haven't been this excited about a network monitoring program since I first came across MRTG about 5 years ago.

The only area I see lacking (if you can call it that) is the fact Cacti is still only just a reporting tool. More often than not sys admins not only require easy to reference and visualise data on infrastructure usage (as Cacti does), but they also need to be alerted to the fact that something is going/gone wrong as system admins are often responsible for monitoring and maintaining 100's or even 1000's of nodes making manual scanning for these events pretty much impossible.

If the devs of Cacti would consider implementing some form of configurable event notification, then I think it would increase it's worth to it's audience greatly. I realise that this is not a small task, and time is money, so even though I'm not rich I'd have no problems donating $100 if this was done. Not enough in itself I know, but it is something.

Would any consideration be given to this?

Khan · Post by **Khan** » Fri Mar 26, 2004 3:10 pm

I second this request. Alerts/Notification are the only missing piece to what is an outstanding piece of software. Keep up the great job!

dieter · Post by **dieter** » Sat Mar 27, 2004 9:36 am

You can write your own alert notification system as I did. Currently supported are banwidth and ping loss & rtt. You write your own datagrid to manage the database which holds the critical values. Then you make a .php file which scans all devices and checks the measured values against the defined critical values. If this is higher then you can send an e-mail, a net send or I even configured to send an SMS during the weekends. You schedule this .php file every 5 minutes.

Khan · Post by **Khan** » Sat Mar 27, 2004 9:52 am

dieter wrote:You can write your own alert notification system as I did. Currently supported are banwidth and ping loss & rtt.

Dieter,
I wish I had the capability to do that but the truth is, I don't. That's why I have to rely on the better coding of real programmers and not my lousy coding. Have you tried sustained high CPU and drivespace alerts? This is something that would come in very handy for those pesky windows machines that I have to monitor.

dieter · Post by **dieter** » Sat Mar 27, 2004 2:57 pm

I just noticed that my program will work for every graph. But you have to enter the critical values manually. So if you have 30 CPU graphs this mean setting 30 times the critical value to 80% (if they all should be considered critical at 80 % off course)

For ping and bandwidth you can set all devices to a specified value by just clicking one button (or still manually every graph if they have different criteria)

Maybe I will post the code on this forum but I'm not sure. I don't want to post a lousy program. First a little more testing...

Wunk · Post by **Wunk** » Sun Apr 04, 2004 3:02 pm

If you could share the code for that, it would be great..

We use cacti to monitor 90+ servers, it lets our customers view their historic graphs of their colo-servers and see the 95th percentile which they've used this month..

Other then that, we use it as a tool to monitor our bandwith, we monitor our uplinks in a custom tree where we can see if anything weird is going on (a hack is often followed by massive bandwith)

Plus we pay for that bandwith

A good add-on for cacti would be that for example:
If a graph gives a value over the threshold for more then a set period of time, a mail could be sent out..

So if a graph with the bandwith would exceed 20 Mbit for more then 15 minutes, an alarm rings so to say, or if the load of a server is higher then 5 for more then an hour, a mail is sent..

This would make cacti double as useful as it is now, especially when everyone and his dog is sleeping at 3am, and something nasty happens

Mika · Post by **Mika** » Fri Apr 16, 2004 8:19 am

I've been thinking about how to make this feature by own.
But before to decide to do so, it would be great to hear Ian's opinion.
Ian, can we expect this feature in near future?
I'm affraid I'm too weak to do so myself

Did · Post by **Did** » Wed Apr 21, 2004 1:33 pm

Hello,
Do you know JFF NMS (http://www.jffnms.org/) ?
This tool is also base on RRD.

It allows to display graph as Cacti, and much more :
- zoom on graphs are possible
- ability to display different reports
- ability to have alarms (and automatic acknoledge in some case)
- alarm on scren or by email,
- display time on graphs (useful when you print them, ...)
- ....

I have discovered when I have studied the opportunity to choose a free tool to make an SLA for the LAN.

It is just a combination of what Cacti and Nagios do.
A great tool.

The only problem for now is the documentation is not yet completed, but people are currently working on it.

Did

achiwork · Post by **achiwork** » Fri May 28, 2004 11:22 am

dieter wrote:I just noticed that my program will work for every graph. But you have to enter the critical values manually. So if you have 30 CPU graphs this mean setting 30 times the critical value to 80% (if they all should be considered critical at 80 % off course)

For ping and bandwidth you can set all devices to a specified value by just clicking one button (or still manually every graph if they have different criteria)

Maybe I will post the code on this forum but I'm not sure. I don't want to post a lousy program. First a little more testing...

I would love you for this.

pestilence · Post by **pestilence** » Wed Jun 02, 2004 7:28 am

I wrotte a small php script which parses the rrd files for some specified filters and if the filters do exist an alarm trigger is sent to a HP Openview NMS system.
It's really easy to send the alarm, the tough part (not that tough) is to writte the script for the parser.

Post by **TheWitness** » Wed Jun 02, 2004 5:13 pm

Please post php script. Thanks, TheWitness

xstream · Post by **xstream** » Wed Jun 09, 2004 2:07 pm

Can someone who created a threshold script, please post on form.
???

Thank you

hakkrdan · Post by **hakkrdan** » Wed Sep 01, 2004 7:31 pm

Did this topic die? I believe that if this functionality was built-in, it would be HIGHLY benificial to the users of Cacti. SNMP possibly taxes the network already on some very involved networks, and it just would not make sense to run another application to do the alerting..

Dieter, as achiwork said, we would love you for this.

Thanks
-dant

Post by **TheWitness** » Wed Sep 01, 2004 7:33 pm

The reason that there is not too much discussion is because it WILL happen. It's just a matter of when.

TheWitness

addlema · Post by **addlema** » Thu Sep 02, 2004 8:22 am

Can someone post an example of their PHP Scripts?? Are you pulling out of the RRD files directly?

Cacti

Alarm/Event based notification

Alarm/Event based notification

A free rrd tool that manages graphs, alerts, allow zoom

Who is online